15:02:27 #startmeeting manila 15:02:28 Meeting started Thu May 7 15:02:27 2020 UTC and is due to finish in 60 minutes. The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:29 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:02:31 The meeting name has been set to 'manila' 15:02:37 o/ 15:02:39 o/ 15:02:41 o/ 15:02:46 hello 15:02:51 courtesy ping: xyang toabctl ganso vkmc amito tbarron 15:02:52 hi 15:02:54 o/ 15:02:56 hi o/ 15:03:23 hello everyone, here's the agenda for today: 15:03:25 #link https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting 15:03:39 hi 15:03:45 #topic Announcements 15:04:10 the final rc for ussuri is due today 15:04:19 #link https://review.opendev.org/#/c/726132/ (Final RC proposal) 15:04:25 #link https://releases.openstack.org/ussuri/schedule.html (Ussuri Release Schedule) 15:05:17 we've not had any significant changes go into the stable/ussuri branch, but this RC catches up to the tip of that branch 15:05:50 please add your votes to that patch, and speak up now if you know any blockers we must consider 15:06:33 next week is release week, and also the first official week into the next cycle 15:07:19 however, if you looked at the bot spam on #openstack-manila, you'll see that we're already proposing many changes targeting the deliverables for victoria 15:08:06 that's today's announcements, anyone has anything else they'd like to share? 15:09:09 lets move on... 15:09:27 #topic OpenStack Ussuri Community Meetings 15:09:31 #link http://lists.openstack.org/pipermail/openstack-discuss/2020-May/014646.html (ML Post on Community Meetings) 15:09:51 There're two virtual community meetings planned for next week 15:10:26 and it'll include a short 5 min presentation on our progress through Ussuri and our plans around Victoria 15:10:44 * vkmc sneaks in 15:10:45 o/ 15:11:15 i'd encourage you to join this to get a pulse around what's happening around other projects as well 15:11:39 hopefully one of the slots is at a convenient time for you to make it 15:12:00 #topic Victoria PTG 15:12:23 Wanted to bring the draft schedule for the upcoming PTG to your attention 15:12:24 #link http://lists.openstack.org/pipermail/openstack-discuss/2020-April/014567.html (ML Post with slots) 15:12:53 As you can see on that email, we've slots for Monday, Tuesday and Friday 15:13:18 these times overlap with cinder, slightly 15:13:24 on tuesday and friday 15:13:46 You can see the rest of the schedule here: 15:13:47 #link https://ethercalc.openstack.org/126u8ek25noy (PTG Slot Signup sheet) 15:13:58 gouthamr: yeah, np at all 15:14:12 we've been assigned the "Cactus" room 15:14:31 🌵 15:14:58 yet to get confirmation on where exactly this room is going to be... but, this looks like the most likely place 15:15:03 #link https://meetpad.opendev.org (Meetpad) 15:15:42 ^ super excited that the folks at opendev-infra are working on a hosted etherpad+jitsi project! 15:16:20 if you have topics that you'd like to bring up, please make sure they're added here: 15:16:23 #link https://etherpad.opendev.org/p/vancouver-ptg-manila-planning (Victoria PTG Planning Etherpad) 15:16:54 i'll groom this, reach out to the owners and add them into rough slots with the times we've reserved 15:17:09 main thing to note is that we break our Tuesday's meeting into two slots 15:18:00 in the hope that our contributors from the east can join one of them 15:19:07 * gouthamr copies lseki's emoji to the etherpad 15:19:23 anything else about the PTG? 15:19:53 #topic Reviews needing attention 15:19:55 lseki++ 15:20:18 We've a new etherpad to track reviews 15:20:22 #link https://etherpad.openstack.org/p/manila-victoria-review-focus 15:20:43 its not necessary to put your code submissions on this etherpad 15:21:23 but, if you're not getting any attention, and would like us to discuss it or find you reviewers, i'll encourage you to add links to your patches there 15:21:34 we'll try and maintain it through the release 15:22:38 ^ reviewers, please take a look at this etherpad and see what's languishing for review and sign up 15:23:08 i wanted to bring up the latter part of that etherpad 15:23:48 we've been adding a number of patches to get our jobs converted to zuulv3 - this is a cycle goal for the victoria cycle 15:24:25 yep 15:24:34 o/ tosky 15:25:06 we've evolved some pretty heavy machinery around our existing "legacy" style jobs - and have bash scripts that do the heavylifting 15:25:24 these are a bit unorthodox and unreliable in the new world 15:25:51 while working on zuulv3, a part of the goal is also to run zuulv3 jobs on stable branches 15:26:01 its a stretch goal according to the original proposal 15:26:23 however, seeing as we maintain our stable branches with such care and commit fixes proactively most of the time 15:26:26 correct: it's not mandatory, but I believe it makes things easier in the long run 15:26:33 ++ 15:26:55 i took tosky's suggestion that we move these job definitions into the manila-tempest-plugin repository 15:27:29 since this repo is branchless, we can go ahead and try to get everything to work there first, and then clean up jobs in the manila repo en-masse 15:27:39 across all stable branches of manila as well 15:28:04 while getting stuff to work, we've had to make a number of fixes to our devstack plugin 15:28:17 these fixes are on the review etherpad at the bottom 15:28:34 yep; at least for the devstack-based jobs, everything should work (== all the changes were backported) from pike onwards 15:29:06 from the grenade side, right now it's just ussuri and master; it will be probably backported to train in a while, when it's proven to be stable 15:29:20 but at this point the bigger work is on the devstack jobs 15:29:31 ah, ack 15:29:46 i haven't taken a look at grenade yet, but that makes sense 15:30:15 i think testing on our stable/pike branch is broken for a while now 15:30:20 same as stable/ocata 15:30:54 hoping we'd get a chance to discuss if we should keep them around, and if anyone cares 15:31:24 i'll add some notes and bring that up at the PTG 15:32:03 but, we've been taking fixes to stable/queens and things are working there, so it'd be good to take these bugfixes to stable/queens at least 15:32:26 so if you see a bunch of backports making devstack plugin changes - this is it 15:32:59 these should be low impact changes that don't affect anyone else, but devstack users 15:33:30 so i'd like to get them in while we labor through job conversions in parallel 15:33:40 any concerns? suggestions? 15:34:37 if you want a big picture view, please consult this board that we're using as personal trackers so far: 15:34:47 #link https://tree.taiga.io/project/gouthampacha-manila-ci-zuul-v3-migration/kanban (taiga kanban board for zuulv3) 15:35:54 vkmc: looks like cephfs-native is ready for review - w00t 15:36:06 oh, another thing while we're here 15:36:08 gouthamr, D: 15:36:10 gouthamr: looks like you are winning :) 15:36:46 dviroel: breaking things fast and early is the key 15:37:00 fast and early = 4+ cycles after zuulv3 debuts :P 15:37:01 I can do that 15:37:07 fixing things... well, that's another thing 15:37:10 iirc there is a weird errors in one of the jobs, but I couldn't find the issue yet, sorry 15:37:29 tosky: yes :| 15:37:32 https://review.opendev.org/#/c/725763/ 15:38:01 ^ tempest config's not getting copied properly causing this one to fail 15:38:30 thanks for looking at it, tosky - my next experiment would be to re-enable glance and nova and see if that's anywhere related to this 15:39:00 i ran that awk script from devstack/inc/meta-config locally, and things looked fine 15:39:51 i had hoped to run a super-minimal job for container and dummy drivers since we don't test data path on either of their test jobs 15:39:57 something is parsed incorrectly, maybe a quoting issue or something else 15:40:52 ack.. i'll ping you outside of this meeting and follow up 15:41:10 "oh, another thing while we're here" 15:41:30 this is also an opportunity to rationalize some of our jobs 15:42:15 we've a "generic-single-backend" job that's functionally identical to "generic-multi-backend" - but from the naming you know the only difference, i doubt its worth keeping two jobs just to test single and multibackend style configurations 15:42:36 maybe they have different settings for the tempest tests? 15:42:39 or it doesn't matter? 15:43:04 slightly different, but they run the same set of tests 15:43:28 slightly different as in, we're spreading testing across multiple backends (same driver) as opposed to a single backend 15:43:40 we also have a "scenario" job that does exactly the same things as the generic driver job, except runs only scenario tests 15:44:33 we have a "generic-no-share-servers" job because we want to test DHSS=False, but that's also being done in LVM, ZFSOnLinux, Dummy and CephFS-{native, nfs} 15:45:05 sigh, so we'll try and chop off some of this excess workload in the check pipeline as necessary 15:45:44 i've had some discussions regarding ^ with dviroel and carloss - especially in light of gate flakiness 15:46:18 +1 manila tests should focus on manila code as much as possible instead of integration testing openstack at tip 15:46:38 time to start penciling in the coverage and see what we can remove and still maintain coverage while conserving test resources and reviewer sanity 15:47:10 not what was ever intended, but most of the failures are due to tip of openstack on master not integrating properly all the time 15:48:24 +1, its a nice goal, but if we have the ability to define periodic jobs too in case we want continuous coverage 15:48:33 s/if// 15:48:56 sorry for going on about this... 15:50:12 let's table any further talk on this, and move on.. 15:50:14 concerns? 15:50:30 please review the bugfix backports :) 15:50:45 ack 15:51:06 #topic Bugs (vhari) 15:51:31 gouthamr, since we are almost out of time .. 15:51:37 here is an easy one :) 15:51:42 #link https://bugs.launchpad.net/manila/+bug/1876820 15:51:42 Launchpad bug 1876820 in OpenStack Shared File Systems Service (Manila) "CI: manila-tempest-dsvm-postgres-container job is failing" [Undecided,In progress] - Assigned to Goutham Pacha Ravi (gouthamr) 15:51:42 hi o/ vhari - yes, sorry about that 15:51:46 nw, 15:52:02 status needs update? in progress 15:52:02 it is an easy one :) 15:52:06 but fix merged 15:52:24 ack, my bad for not setting those fields 15:52:27 checking in case I missed something 15:52:40 cool .. 15:52:46 gouthamr, ty 15:52:46 this is a high, because it broke the container job 15:53:20 but, it's being worked around and i'll create another bug to address the main issue 15:54:19 done 15:54:31 :) that's a wrap for bugs 15:55:05 great thank you for your patience and effort with the triage vhari!! 15:55:10 #topic Open Discussion 15:55:19 we've all of five minutes 15:55:21 gouthamr, yw 15:55:28 lets talk some more zuulv3? 15:55:50 okay no, i was kidding :) 15:56:07 is zuulv4 out yet? 15:56:39 tbarron: might skip that version and ship zuul5 :) 15:57:47 alright everyone, thank you so much for joining. I hope everyone continues to keep safe and healthy 15:57:58 see you all on #openstack-manila 15:58:11 #endmeeting